Vibe Wrapper

Local-first spike for reading vibe coding session data from Codex, Claude Code, Cursor, and TokenTracker activity buckets.

Inspect Local Sessions

npm run inspect -- --from 2026-06-01 --to 2026-06-08 --sources codex,claude,cursor

The command prints JSON with source counts, prompt counts, token totals where available, TokenTracker-backed activity rows when available, word frequencies, and prompt records. It also includes profile_signals, which separates useful intent prompts from pasted code/log reference material and reports local Codex setup signals such as skill count, MCP servers, enabled plugins, and custom instructions. Reference material is still analyzed for code/log signals such as languages, file extensions, file paths, and error types.

Preview Profile Signals

node - <<'NODE'
const { inspectSources } = require('./src/inspect');
(async () => {
  const report = await inspectSources({
    from: '2026-06-01',
    to: '2026-06-08',
    sources: ['codex', 'claude', 'cursor'],
  });
  const prompt = report.profile_signals.prompt_analysis;
  const env = report.profile_signals.environment.codex;
  console.log(JSON.stringify({
    summary: report.summary,
    prompt_counts: {
      total: prompt.total_prompts,
      useful: prompt.useful_prompt_count,
      reference: prompt.reference_prompt_count,
      useful_ratio: prompt.useful_ratio,
    },
    categories: Object.fromEntries(
      Object.entries(prompt.categories)
        .map(([name, row]) => [name, row.count])
        .filter(([, count]) => count > 0)
    ),
    reference_signals: prompt.reference_summary.signals,
    environment: {
      skills: {
        total: env.skills.count,
        user: env.skills.user_count,
        plugin: env.skills.plugin_count,
        sample: env.skills.names.slice(0, 12),
      },
      mcp_servers: env.mcp_servers,
      plugins_enabled: env.plugins.enabled_count,
      custom_instructions_chars: env.custom_instructions.char_count,
    },
    top_terms: report.word_frequencies.slice(0, 15),
  }, null, 2));
})();
NODE

Default local sources:

• Codex: ~/.codex/sessions
• Claude Code: ~/.claude/projects
• Cursor: ~/Library/Application Support/Cursor/User/globalStorage/state.vscdb on macOS
• TokenTracker activity: ~/.tokentracker/tracker/queue.jsonl

Cursor support is currently best-effort. It reads user bubbleId rows from cursorDiskKV, but those compact prompt rows may not include reliable timestamps or token usage.

The dashboard 3D heatmap uses TokenTracker hourly token buckets when that queue exists. If it is missing, the heatmap falls back to prompt counts from the prompt adapters.